Why Meta will update AMM policy

How does this help with advertising?
How does this help with iOS advertising?
What are the benefits of Facebook doing this?

First of all, they will definitely not do something that is not profitable. They used to not provide data to everyone under the banner of privacy, but now they suddenly don’t install it? What is the benefit? In order to get more advertising budget, many products’ BI designs are based on device-level data after people could not get device-level data. If you can’t get device-level data, there will be a lot of restricted data in the original data (some data can be obtained through referral, but there is always a 10-30% range that cannot be obtained. iOS is even worse. You may get a little more after AEM). In this way, the BI made by customers will have less data and cannot truly judge the LTV of users in this channel. This will make some customers feel that FB’s effect is not good and reduce their budget. This year, the mutual volume of advertising platforms has also become serious. Even Google finally couldn’t stand it and started to do “probabilistic attribution”. I forgot what name they gave it, but the essence is that they couldn’t stand it and simply worked with third parties to do probabilistic attribution by themselves, so that customers can see similar data from accounts and third parties. (Google’s effect on iOS was a complete mystery before, causing many advertisers to simply not run Google’s iOS ads.) After knowing the platform’s benefits, it is clear what kind of help it will have on the delivery side. In order to get more data for us, the platform’s own BI can analyze more accurate conversion effects, so as to allocate more budget to Meta. 1. It is better to do internal LTV evaluation, and BI design can also be more accurate. 2. In theory, the Meta backend will change the model strategy. It used to rely more on SKAN data, but now it may use probabilistic attribution to improve advertising accuracy.

Advertising platforms are also increasing their efforts on the “accuracy” of probabilistic attribution, so as to make third-party data and account data closer, in order to obtain more budget.